Utility of non-invasive sensors for the prediction of heart failure deterioration after recent hospitalization due to acute decompensated heart failure: protocol for the RM4-HF observational trial

Abstract Background Remote patient monitoring has been proposed as an effective strategy to reduce hospitalizations and mortality in patients with heart failure. Current remote patient monitoring approaches typically rely on spot measurements of parameters such as weight, heart rate and blood pressure. Recent advances in non-invasive sensing technologies have enabled continuous monitoring of cardiovascular parameters, offering an more comprehensive view of patients’ cardiovascular status. These sensors may enhance remote patient monitoring strategies by enabling earlier detection of heart failure deterioration. This study aims to evaluate the added value of multiple non-invasive sensors for predicting heart failure deterioration using multiparametric predictive modeling. Methods Patients admitted to the hosptial for acute decompensated heart failure, regardless of type or etiology, will be enrolled at the time of discharge. Upon returning home, digitally literate participants will begin remote patient monitoring combined with two non-invasive sensing techniques. They will continuously wear a wrist-worn data logger equipped with photoplethysmography and accelerometry sensors and will perform daily voice recordings using the ListenHF mobile application. During a three-month follow-up period, continuous sensor data and daily voice samples will be collected. In the event of heart failure deterioration within this period, sensor and remote patient monitoring data will be analyzed using artificial intelligende-based predictive modeling to evaluate whether the model could have predicted the deterioration in advance. Discussion Recent meta-analyses have demonstrated that remote patient monitoring can reduce heart failure hospitalizations and mortality. However, considerable heterogeneity exists among current remote patient monitoring strategies, and technological innovation remains limited. Conventional parameters such as weight, heart rate and blood pressure often change late in the cascade of heart failure deterioration. We hypothesize that incorporating novel non-invasive sensors enables earlier detection of clinical worsening. The collected parameters will be retrospectively analyzed using artificial intelligende to develop a multiparametric predictive model for early identification of heart failure decompensation. Conclusion This study will provide insights into the potential of combining continuous cardiovascular sensing and speech analysis with conventional remote patient monitoring to enable earlier detection of heart failure deterioration. By evaluating the added predictive value of these non-invasive technologies, the findings may support the development of more proactive, personalized prevention strategies aimed at reducing heart failure–related hospitalizations and improving clinical outcomes.Study design diagramFor image description, please refer to the figure legend and surrounding text.